ECTS OF CHANGES OF BIRTH AND DEATH RATES 641
ng one. Moreover, studies in Taiwan suggest that the major increase in use
irth control is at the older ages.
?he results so far are based on a one-sex model and strictly apply only to
female component of population. A more precise analysis would take )unt of males explicitly. It turns out, for example, that males are rela-ly few in the United States today, at the ages at which they become .ers, in comparison with the number of females at the ages at which they Dme mothers. This observation applies to other countries affected by the twar baby boom, [he disparity has been called the marriage squeeze. We do not know what
of adjustment is made by those of marrying age to the fewness or surplus suitable marriage partners. With sharp drops in birth rates anticipated, sequences such as this call for more research.
ADEQUACY OF DEMOGRAPHIC DATA
Despite all the effort that has been made, present-day statistics are far n complete; for only about 30 percent of the world population do cen-ss and vital registrations exist that provide a minimal basis for demo-)hic analysis. Fortunately, this coverage includes the three kinds of ulation as classified by level of mortality and fertility: high birth and high th rates, high birth and low death rates, and low birth and low death rates, remaining configuration, low birth and high death rates, is not of usual urrence.
Statistical information is far from equally available for the three groups, pies of low birth and low death rates are also those that are most devel-d economically, and since statistical advance tends to accompany other ctions of advance, the existing data overrepresent developed countries. 3 is particularly true of birth and death registrations, dependent as they on the education of the public as a whole, as well as of statistical person-Machinery for censuses, on the other hand, can be set up more quickly, a much larger part of the world's population has recently been counted in suses than has had its births and deaths registered.
Che opposite was true of Europe in the 18th century when baptismal and ial records were widespread and often of high quality, but national organ-ions for census-taking were not yet in being.
foday census-taking is ahead of vital registration not only in extent, but
> in completeness. For a census to be more than 20 percent short is rare,
published birth figures sometimes appear to include no more than one
' of the births that occurred. Fortunately the relation of births and census
,-! 1'o +-4 U,,+;.-,.-,,-, +r. U« +r,1^.-i ,,* lot^.- Q^nKlao oo-f ;,•>-,<> too r\f Kit-tin fjfao tr\ lio. Because some women die after the age of 20 and before 40, the reduction of the birth rate at the younger age will have more consequence for the rate of natural increase. For a population that is rapidly increasing, however, this is lessh to prove troublesome. control. Dr. Hilton Salhanick has observed that some women practicing the rhythm method will break or lose their thermometers at the critical juncture in theirright, therefore alwayshas some c